1 |
Why not call a spade a spade? As a user I think *kernel* version |
2 |
numbers are good - you know exactly what you are getting. I do agree |
3 |
that in general, using incremental version numbers is bad, but Kernels |
4 |
are a special case. What happens when we get to 2.8? |
5 |
gentoo-dev-dev-sources? We already use modules.autoload.d/kernel-2.4 |
6 |
and 2.6 because the differences are so great - this is quite a neat fix |
7 |
by the way. Also linux26-headers which is what they truly are (just |
8 |
wish they and emerge would play nicely together) |
9 |
|
10 |
There is a lot of confusion in userland by not calling kernels by their |
11 |
real name - I include 2.4 and 2.6 as part of the kernel name (i.e., as |
12 |
alpha text), and the sub version as the true version. How many times do |
13 |
you see something like "I am running a 2.6 system" - nobody says "I am |
14 |
running a gentoo-dev-sources system" as outside of gentoo (and possibly |
15 |
limited even there), users would have no idea what kernel version that |
16 |
is - and this information is critical as to how a system works. |
17 |
Kernel.org can get away with calling 2.6 current, but a distro does have |
18 |
to take into account legacy in a sane, manageable and user friendly |
19 |
fashion. |
20 |
|
21 |
|
22 |
|
23 |
BillK |
24 |
|
25 |
|
26 |
On Wed, 2004-08-25 at 23:42, Cory Visi wrote: |
27 |
> On Tue, Aug 24, 2004 at 09:26:06AM -0400, Chris Gianelloni wrote: |
28 |
|
29 |
> 2.6 as the "default" kernels in the near (February) future. |
30 |
> |
31 |
> This is still using version numbers in package names. I do not think we |
32 |
> should take this approach. Aside from modifying portage a little, I liked |
33 |
> the "legacy-sources" approach the best. |
34 |
> |
35 |
> -Cory |
36 |
> |
37 |
> -- |
38 |
> gentoo-dev@g.o mailing list |
39 |
> |
40 |
> |
41 |
|
42 |
|
43 |
-- |
44 |
gentoo-dev@g.o mailing list |